JP2018131071A - System and method for object detection - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system and a method for object detection which is easy to control and maintain and has high reliability.SOLUTION: An object detection system 7 comprises a distance image sensor 71 which measures an object area and a control part 72 enabling the sensor 71 to acquire background data and the control part 72 checks the operation state of the sensor 71 by comparing the first background data with the second background data under the condition that these data are acquired at different time.SELECTED DRAWING: Figure 3

Description

  The present invention relates to an object detection system and an object detection method used on a platform or the like.

  As an object detection system or method for a platform, one that generates a boundary of a detection area by adjusting an offset with respect to the distance of each pixel in a three-dimensional distance image acquired by a distance image sensor is known (Patent Document). 1). In this system or method, when setting the boundary of the detection area, when a predetermined ratio or more of pixels cannot be measured on the trajectory side of the three-dimensional distance image, it is determined that the vehicle is absent, and the trajectory side of the three-dimensional distance image When it is impossible to measure less than a predetermined ratio of pixels, it is determined that a vehicle is present.

  Further, it is a movable home gate device for a platform, and the area between the outside of the fence and the train that stops on the platform is included in the detection area, and the detection area of the detection device is set for each train that stops on the platform, or For each vehicle that forms a train formation, it is known to update the vehicle according to the position and shape of the side surface of the vehicle (Patent Document 2).

  However, in the systems or apparatuses described in Patent Documents 1 and 2, it is assumed that each unit such as the distance image sensor operates normally, and the system itself predicts deterioration of the distance image sensor and the progress thereof. It is not easy to determine the degree of management and maintenance suitable for a system that requires high reliability.

Japanese Unexamined Patent Publication No. 2017-19380 JP2013-203118A

  The present invention has been made in view of the above-described background art, and an object thereof is to provide an object detection system and an object detection method that are easy to manage and maintain and have high reliability.

  In order to achieve the above object, an object detection system according to the present invention includes an image sensor that performs measurement on a target region and a control unit that causes the image sensor to acquire background data. The operating state of the image sensor is checked by comparing different first background data and second background data.

  According to the object detection system, the control unit checks the operation state of the image sensor by comparing the first background data and the second background data, which are different in acquisition timing. The health of the image sensor can be judged by using the background data acquired at multiple timings that are set appropriately as in the case of performing an object, and the management and maintenance of the object detection system is facilitated and its reliability is increased. be able to.

  According to a specific aspect of the present invention, in the object detection system, the image sensor is a distance image sensor that acquires distance data for a target region, and the control unit transmits background data to the distance image sensor when no object is detected. To get. In this case, the soundness of the image sensor can be determined using background data measured for background removal by the distance image sensor.

  According to another aspect of the present invention, the background data is acquired for the target region including the track side of the platform door by the distance image sensor at the timing when the train image is detected by the distance image sensor at the time of non-detection. In this case, the soundness of the distance image sensor or the like can be determined using the background data acquired and updated according to the train line.

  According to still another aspect of the present invention, the control unit obtains background data for each train entry, and obtains the first background data obtained at the time of arrival of the current train at the time of arrival of the previous train at the second time. Compare with background data. In this case, background data can be updated for each train line, and a detailed check of the operation state is possible.

  According to still another aspect of the present invention, the control unit determines that the state is abnormal when the difference value or the detection pixel number difference between the first background data and the second background data exceeds a predetermined limit. To do.

  According to still another aspect of the present invention, the control unit determines that the state is abnormal when the difference value exceeds a predetermined limit set based on an assumed upper limit value of the temperature change. In this case, it is possible to prevent the change with time due to the temperature change from affecting the evaluation of the operation state, and it is possible to improve the reliability of the soundness check of the object detection system.

  According to still another aspect of the present invention, the control unit causes the distance image sensor to acquire background data when detecting the train entry based on the output of the distance image sensor. In this case, the train entry can be determined without depending on the external system, so that the object detection system can be easily installed.

  According to still another aspect of the present invention, the control unit detects an object as a difference between background data and distance data detected by a distance image sensor at the time of detection. In this case, an obstacle or other object can be detected as a three-dimensional portion protruding from the background data.

  According to still another aspect of the present invention, the control unit corrects the background data by performing a predetermined offset adjustment on the background data. In this case, even if the characteristics of the distance image sensor fluctuate somewhat due to the influence of temperature or the like, it is possible to suppress malfunction.

  According to still another aspect of the present invention, the control unit performs offset adjustment for each pixel or for each block including a plurality of pixels.

  According to still another aspect of the present invention, the image sensor is provided in each of a plurality of platform doors provided along the platform end. In this case, it is possible to reliably determine whether a person or an object remains between the platform door and the track or the vehicle, and to ensure proper operation of the platform door and safe operation of the vehicle.

  In order to achieve the above object, an object detection method according to the present invention is an object detection method in which an object is detected by measuring a target region using an image sensor, and is obtained by a first timing acquired by the image sensor at different timings. The operating state of the image sensor is checked by comparing the background data and the second background data.

  According to the object detection method, since the operation state of the image sensor is checked by comparing the first background data and the second background data, which are different in acquisition timing, for example, when measuring the target region As described above, it is possible to determine the soundness of the image sensor by using the background data acquired at a plurality of timings appropriately set, and the management and maintenance of the object detection system can be facilitated and the reliability thereof can be improved.

It is the schematic explaining the management system containing the object detection system of embodiment. It is a notional perspective view explaining a home door device and an object detection system. It is a block diagram explaining an object detection system. It is a figure explaining an example of the distance image sensor integrated in an object detection system. (A) is a plan view of background data, (B) is a side view of background data, and (C) is a side view of background data in a different scene corresponding to (B). It is a figure explaining offset adjustment to background data. It is a conceptual diagram explaining operation | movement of the object detection system shown in FIG.

  Hereinafter, an embodiment of an object detection system according to the present invention will be described with reference to the drawings.

  FIG. 1 is a conceptual diagram illustrating a home door system incorporating an object detection system. The home door system 300 includes a home installation unit 100, an overall control unit 200, and a communication network 9.

  The home installation unit 100 includes a number of platform door devices 3 arranged along the platform end 2 a of the platform 2 and an object detection system 7 provided along with each platform door device 3. The overall control unit 200 performs overall management of operations of the platform door device 3 and the like. The communication network 9 connects the home door device 3 and the object detection system 7 constituting the home installation unit 100 in a communicable manner.

  As shown in FIG. 2, the platform door device 3 is provided corresponding to each of the vehicle doors TD that are provided in a plurality of one or more trains TR that are connected to the platform end 2a and stop at a specified position. It has been. The home door device 3 includes a pair of partition walls 4 that partition the inner and outer regions 2b and 2c of the platform 2, and a pair of home doors 6 that are openably and closably installed at positions facing the vehicle door TD. Note that the partition wall 4 may also be used as the adjacent platform door device 3.

  The object detection system 7 is installed between the partition wall 4 of each platform door device 3 and the vehicle 1, and is fixed to the vehicle 1 side of the partition wall 4. The object detection system 7 monitors the outer region 2b on the track side of the platform door device 3, and detects the state of the platform door 6 on the train TR side and its surroundings. Specifically, the object detection system 7 is an obstacle detection system for the platform door device 3, and has a role of detecting obstacles remaining on the train TR side of the platform door 6 or in the vicinity thereof. In the example shown in FIG. 2, the object detection system 7 is attached only to one cut wall 4 of the pair of home doors 6, but the object detection system 7 may be provided on both sides of the pair of home doors 6. Moreover, it is not necessary to provide the object detection system 7 for each platform door device 3, and the object detection system 7 may be provided in units of a plurality of platform door devices 3.

  As shown in FIG. 3, the object detection system 7 includes a distance image sensor 71, a control unit 72, and a communication unit 73. As will be described later, the distance image sensor 71 is an image sensor that acquires distance data for the obstacle OS in the target region by using the laser beam LB. The control unit 72 causes the distance image sensor (image sensor) 71 to acquire background data and distance data (that is, distance image data). The control unit 72 calculates an object detection image as a difference between the distance data and the background data obtained by the distance image sensor 71. Further, the control unit 72 converts the object detection image and the distance data into information required by the external device (for example, the distance data colored according to the distance), or the control information from the external device. The distance image sensor 71 is received and the operation state of the distance image sensor 71 is adjusted. The control unit 72 includes a storage unit 72a that stores programs, data, and the like necessary for processing. The communication unit 73 connects the object detection system 7 to the overall control unit 200 via the communication network 9 (see FIG. 1) so as to be communicable, and connects the plurality of object detection systems 7 via the communication network 9 so as to be communicable. To do.

  Returning to FIG. 1, the overall control unit 200 is a device that is operated by a station staff and incorporates a computer or the like. The overall control unit 200 can monitor and remotely control the operating states of the platform door device 3, the object detection system 7, and the like.

  The structure of the distance image sensor (image sensor) 71 shown in FIG. 3 will be described with reference to FIG. The distance image sensor 71 includes a two-dimensional scanner 10 having a movable mirror 11a that can be swung in two axial directions. In the present embodiment, a planar actuator is used as the two-dimensional scanner 10 ( JP, 2011-033980, A, etc.). In addition to the planar actuator, a polygon mirror, a galvanometer mirror, or the like may be used.

  In addition to the two-dimensional scanner 10, the distance image sensor 71 includes a sensor control unit 20, a laser projector 30, a laser receiver 40, a distance measuring unit 50, and the like.

  The sensor control unit 20 is provided with a laser controller 21 for controlling operations of the two-dimensional scanner 10 and the laser projector 30. The laser controller 21 causes the scanner driver 22 to output a drive signal to the two-dimensional scanner 10 and receives a scanner synchronization signal from a detection unit (not shown) of the two-dimensional scanner 10 via the filter 23.

  In the laser projector 30, the laser driver 31 operates the laser element 32 based on the emission timing signal from the laser controller 21. Laser light L1 emitted from the laser element 32 is projected onto the two-dimensional scanner 10 via the light projecting optical unit 33 and the beam splitter 61. The light emission monitor 38 monitors the light projection timing by the light projection optical unit 33 by partially branching and monitoring the laser light L1 projected from the light projection optical unit 33.

  In the laser light receiving unit 40, the light receiving element 41 detects the laser light L <b> 2 reflected by the two-dimensional scanner 10 and returned via the beam splitter 61 via the light receiving optical unit 42. The preamplifier 43 amplifies the light reception signal from the light receiving element 41, and the A / D converter 44 A / D converts the light reception signal output from the preamplifier 43, and calculates the distance value in the sensor control unit 20 as the light reception amount. Output to the circuit 25.

  The distance measuring unit 50 detects the light reception signal sent from the preamplifier 43 of the laser light receiving unit 40, and therefore, the resonance circuit 51 used mainly when the distance to be measured is relatively long and the distance to be measured are relatively short. A rising circuit 52 is provided for use in some cases. In the subsequent stage of the resonance circuit 51 and the rising circuit 52, stop timing generation circuits 54 and 54 that generate timings for receiving the returned laser beam, and a time measurement circuit 55 that measures the time from the light projection timing to the light reception timing. , 55 and A / D converters 56, 56 for outputting time data.

  In the sensor control unit 20, the distance value calculation circuit 25 receives time data generated by the time measurement circuits 55 and 55 and the A / D converters 56 and 56 of the distance measurement measurement unit 50. The distance value calculation circuit 25 is configured to be able to acquire a distance value based on the time data from the time measurement circuit 55 and the like, and the acquired distance value is shown in FIG. It is output to the control unit 72.

  The sensor control unit 20 is provided with a main power supply 29 for supplying power to the circuit portion in the sensor control unit 20, the distance measurement measurement unit 50, and the HV power supply 63. The HV power supply 63 that receives the power is a driving power supply for the laser driver 31, the preamplifier 43, and the scanner driver 22.

  The outline of the operation of the distance image sensor 71 will be described. While the movable mirror 11a of the two-dimensional scanner 10 is periodically swung in two orthogonal directions, the laser light L1 is emitted from the laser projector 30 to the two-dimensional scanner 10. Is caused to enter, and the laser beam LB is scanned in a two-dimensional region. The laser light L2 reflected by the two-dimensional scanner 10 is reflected by the two-dimensional scanner 10 and reflected by the obstacle OS (refer to FIG. 3). Distance data or distance values can be acquired at each pixel constituting the target area that is the range scanned by the dimension scanner 10.

  With reference to FIG. 5 (A) and 5 (B), the detection area by the distance image sensor 71 is demonstrated. The detection area AD by the distance image sensor 71 corresponds to a measurement target area by the distance image sensor 71, and is set within a maximum range in which the laser beam LB can be projected. The specific detection area AD is the space above the outer region 2c of the platform end 2a sandwiched between the platform door 6 or the outer wall 4d of the partition wall 4 and the vehicle door TD or the vehicle outer wall TW. It has become. The specific detection area AD covers the front surface of the vehicle door TD and the vicinity thereof, and is set across the outer walls 4d of the pair of adjacent partition walls 4. As a result, the detection area AD is a rectangular parallelepiped space, and has a width D1 corresponding to the distance between the outer wall 4d and the vehicle door TD in the Y direction which is perpendicular to the X direction in which the platform end 2a extends and is horizontal. In the X direction in which the platform end 2a extends, the width D2 is sufficiently wider than the interval between the partition walls 4, and in the Z direction perpendicular to the X direction in which the platform end 2a extends, the height of the vehicle door TD is high. It has a moderate width D3.

  The boundary ADa on the vehicle door TD side of the detection area AD corresponds to the vehicle door TD and the vehicle outer wall TW, and is obtained as background data by detecting the vehicle door TD and the vehicle outer wall TW by the distance image sensor 71. The boundary ADb of the detection area AD on the platform door device 3 side corresponds to the platform door 6 or the outer wall 4d, and is obtained as background data by detecting the platform door 6 or the outer wall 4d by the distance image sensor 71. . The boundary ADc on the platform end 2 a side of the detection area AD is obtained as background data by detecting the outer region 2 c of the platform end 2 a by the distance image sensor 71. The backgrounds ADd and ADe on the upper side and the back side of the detection area AD have a background that is quite far away, and therefore virtual background data that is forcibly set is used. However, the actual background can also be used for the boundaries ADd and ADe. Note that the boundary ADf on the near side of the detection area AD is defined by the scanning range of the laser beam LB by the distance image sensor 71.

  In the above, the detection area AD when operating the distance image sensor 71 at the timing when the train TR is detected is described. However, the detection area AD ′ at the timing before the train TR is detected, that is, the train TR The detection area AD ′ when the distance image sensor 71 is operated to detect an incoming line is open to the track side where the train TR comes. That is, as shown in FIG. 5 (C), the line-side boundary ADa ′ of the detection area AD ′ is made to recede to the line side by setting the line-side boundary ADa of the detection area AD to a predetermined distance or more. It is said. The boundary ADa 'can be a background far away, but can also be a virtual background data that is forcibly set. The detection area AD ′ shown in FIG. 5C can be used when detecting an incoming line of the train TR. That is, the object detection system 7 periodically or periodically determines whether or not the train TR has entered the line when no object or obstacle is detected. When determining whether or not there is an incoming line, if the train TR enters, the distance image sensor 71 acquires distance data corresponding to the detection area AD shown in FIG. The train TR is detected as a difference between the distance data most recently detected by the sensor 71 and the background data shown in FIG. If there are a predetermined number or more of pixels having a difference between the distance data and the background data equal to or greater than a predetermined value, the control unit 72 determines that the train TR has been entered.

  In addition, after the vehicle door TD and the platform door 6 are opened and the user gets on and off, when the vehicle door TD is closed and the user gets on and off, the object detection system 7 detects the object or obstacle OS. Works as a detection. In this case, the control unit 72 detects the object, that is, the obstacle OS, as a difference between the distance data detected most recently by the distance image sensor 71 and the background data shown in FIGS. 5 (A) and 5 (B). The control unit 72 determines that the obstacle OS remains in front of or around the vehicle door TD if there are a predetermined number or more of pixels in which the difference between the distance data and the background data is a predetermined value or more.

  FIG. 6A is a diagram for explaining correction of background data. When the background surface 1a is scanned with the laser beam LB from the distance image sensor 71 to obtain background data, it is desirable to perform offset adjustment. The distance data DB of the surface 1a obtained by the distance image sensor 71 is substantially along the surface 1a. If this distance data DB is used as background data as it is, environmental conditions such as the temperature at which the distance image sensor 71 is placed. Therefore, noise is easily generated when the obstacle OS is detected thereafter. For this reason, a value obtained by subtracting a predetermined distance from the distance of the original background data or distance data DB is defined as a detection area AD, that is, new corrected background data DD. In a specific method, pixel distance data indicating a three-dimensional boundary distance is obtained as the original background data or distance data DB. Therefore, an offset amount, for example, several cm to several Set 10 cm. That is, a new value obtained by subtracting the offset amount from each pixel distance data constituting the original background data or distance data DB is a new pixel distance data after correction, thereby providing a new margin with noise. Background data DD or detection area AD can be obtained.

  In the above description, the offset amount is adjusted or set in units of pixels. However, the offset amount can be adjusted or set in units of blocks including a plurality of pixels. By using the block unit, the burden of calculation processing in the control unit 72 can be reduced. The offset amount can be set individually for each of the boundaries ADa, ADb, and ADc of the detection area AD. In addition, a plurality of regions having different offset amounts can be provided in the boundaries ADa, ADb, and ADc. Similarly, the number of pixels included in the block can be individually set for each of the boundaries ADa, ADb, and ADc of the detection area AD. From the plurality of types of blocks having different numbers of pixels in the boundaries ADa, ADb, and ADc. Can be provided.

  The operation of the object detection system 7 will be described with reference to FIG. First, the control unit 72 performs object detection that calculates a difference from distance data obtained by the distance image sensor 71 with reference to the background data shown in FIG. 5C when no object or obstacle OS is detected. Then, it is determined whether or not the train TR has entered the platform 2 (step S11). When the distance image or the object detection image of the vehicle door TD shown in FIG. 5A (corresponding to the boundary ADa) or the object detection image is obtained (YES in step S11), the control unit 72 obtains the distance data obtained by the distance image sensor 71. The background data is acquired from (step S12). The background data is not limited to the original background data, and may be data obtained by performing processing such as offset adjustment on the background data. The processed data is also background data in a broad sense. In this case, since the train TR exists on the track side of the platform end 2a, background data (first background data) AD as shown in FIGS. 5 (A) and 5 (B) is obtained. The first background data AD1 is temporarily stored in the storage unit 72a. Next, the control unit 72 reads the background data (second background data) AD acquired at the time of the previous entry of the train TR from the storage unit 72a (step S13). FIG. 6B is a conceptual diagram comparing the first background data AD1 and the second background data AD2. Next, the control unit 72 takes the difference between the first background data AD1 and the second background data AD2 (step S14). When the difference value between the first background data AD1 and the second background data AD2 exceeds a predetermined limit, the control unit 72 determines that the operation state of the distance image sensor 71 is an abnormal state. As the difference value, for example, various index values such as the maximum distance difference for each pixel and the root mean square of the distance difference can be used. When it is determined that the difference value, that is, the background data difference exceeds the predetermined limit and is outside the allowable range (YES in step S14), the control unit 72 determines that the operation state of the distance image sensor 71 is an abnormal state and causes an error. A signal is output (step S15). Here, the control unit 72 detects the distance image sensor when the difference between the first background data AD1 and the second background data AD2 exceeds a predetermined limit set based on an upper limit value of the assumed temperature change. It is determined that the operation of 71 is in an abnormal state. As a result, it is possible to prevent time-dependent fluctuations associated with temperature changes from affecting the evaluation of the operating state, and to improve the reliability of the soundness check of the object detection system 7. An error signal from the control unit 72 is output to the overall control unit 200 via the communication network 9. Since the station staff who operates the overall control unit 200 receives information or a warning that the background data of the distance image sensor 71 indicates an abnormality via a display (not shown) or the like provided in the overall control unit 200, the distance image It is possible to quickly detect the deterioration of the sensor 71 and its sign. The error signal includes not only information regarding whether or not the index value obtained from the difference between the first background data AD1 and the second background data AD2 exceeds a predetermined reference value, but also indicates which index value the predetermined reference value. It can include whether or not the degree is exceeded. In this case, the station staff can also determine the degree of progress of the deterioration of the distance image sensor 71 from the display of the overall control unit 200 or the like.

  As a result of comparing the difference between the first background data AD1 and the second background data AD2, when the background data difference is determined to be within the allowable range (NO in step S14), the control unit 72 The offset processing illustrated in FIG. 6A is performed on the data, that is, the first background data AD1 obtained in step S12, and the first background data AD1 after the offset processing is used as the background data AD ( Step S21). Next, the control unit 72 sets that it is a detection time when the object detection system 7 detects an obstacle, and acquires distance data by the distance image sensor 71 (step S22). Next, the control unit 72 detects an object, that is, an obstacle OS as a difference between the distance data obtained in step S22 and the background data AD obtained in step S21 (step S23). When the obstacle OS is detected (YES in step S23), the control unit 72 outputs an obstacle signal, that is, reports (step S24), returns to step S22, and repeats the processes of steps S22 to S24. If no obstacle OS is detected (NO in step S23), the control unit 72 determines that no obstacle remains in front of or around the vehicle door TD, and controls the overall control unit 200 via the communication network 9. The obstacle absence signal is output to (step S25). In this case, the station staff can confirm a display permitting the closing of the platform door 6 via the display of the overall control unit 200 or the like.

  As described above, in the present embodiment, the sensor control unit 20 compares the first background data AD1 and the second background data AD2 that have different acquisition timings, thereby obtaining a distance image sensor (image sensor). Since the operation state of 71 is checked, for example, the soundness of the distance image sensor (image sensor) 71 is determined by utilizing the background data AD acquired at a plurality of timings set appropriately as in measurement of the target region. Therefore, the management and maintenance of the object detection system 7 can be facilitated and its reliability can be improved.

  In the object detection system 7, the distance image sensor (image sensor) 71 acquires distance data for the target region, and the sensor control unit 20 detects the distance image sensor (image sensor) when no object or obstacle OS is detected. 71 is made to acquire background data. In this case, the soundness of the image sensor (image sensor) 71 can be determined using background data measured for background removal by the distance image sensor (image sensor) 71.

  The home door system 300 includes an object detection system 7 for each home door device 3, and a plurality of object detection systems 7 arranged along the platform 2 form a complex object detection system. . That is, the distance image sensor (image sensor) 71 is provided in each of the plurality of home doors 6 provided along the platform end 2a. In this case, it can be reliably determined whether or not people or objects remain between the platform door 6 and the track or the train TR, and proper operation of the platform door 6 and safe operation of the vehicle can be ensured. .

  The structure, shape, size, and arrangement relationship described in the above embodiments are merely schematically shown to the extent that the present invention can be understood and implemented. Therefore, the present invention is not limited to the described embodiments, and various modifications can be made without departing from the scope of the technical idea shown in the claims.

  Although the case where the distance image sensor 71 is used as an image sensor has been described in the above embodiment, a camera that captures an object with a wavelength in the infrared or visible region can be used as the image sensor. In this case, the obstacle OS and other objects are detected as differences in luminance and pattern from the background data obtained from the background image.

  In the above embodiment, when the difference between the first background data AD1 and the second background data AD2 exceeds a predetermined limit set based on the assumed upper limit value of the temperature change, the control unit 72 performs the distance image. Although it is determined that the operation of the sensor 71 is in an abnormal state, the difference between the number of detected pixels of the first background data AD1 and the number of detected pixels of the second background data AD2, that is, detection of the background data AD1, AD2 is detected. When the difference in the number of pixels exceeds a predetermined limit, a process for determining that the operation of the distance image sensor 71 is in an abnormal state is also possible.

  In the above embodiment, the difference between the first background data AD1 obtained at the current timing when the train TR enters the platform 2 and the second background data AD2 obtained at the previous timing when the train TR entered before. Although the value is calculated, the second background data AD2 can also be background data obtained at the timing when the train TR has entered the line twice or more before.

  In the above embodiment, the background data AD1 and AD2 are measured at the timing when the train TR enters the platform 2, but the train TR enters the background data AD1 and AD2 for obtaining a difference value or a difference in the number of detected pixels. Not only the timing but also a regular timing such as a specific time of the day when the train TR is not connected.

  DESCRIPTION OF SYMBOLS 2 ... Platform, 2a ... Platform edge, 2b, 2c ... Inner and outer area, 3 ... Home door apparatus, 4 ... Wall, 4d ... Outer wall, 6 ... Home door, 7 ... Object detection system, 9 ... Communication network, 10 ... Two Dimensional scanner, 11a ... movable mirror, 20 ... sensor control unit, 30 ... laser projection unit, 40 ... laser receiving unit, 50 ... distance measurement measurement unit, 71 ... distance image sensor, 72 ... control unit, 73 ... communication unit, DESCRIPTION OF SYMBOLS 100 ... Home installation part, 200 ... General control part, 300 ... Home door system, AD ... Background data, AD1 ... First background data, AD2 ... Second background data, ADa, ADb, ADc, ADd, ADe, ADf ... Boundary, DB ... Distance data, DD ... Background data, L1, L2, LB ... Laser light, OS ... Obstacle, TD ... Vehicle door, TR ... Train, TW ... Vehicle outer wall

Claims (12)

An image sensor for measuring the target area;
A control unit that causes the image sensor to acquire background data,
The said control part is an object detection system which checks the operation state of the said image sensor by comparing the 1st background data and 2nd background data from which acquisition timing differs mutually. The image sensor is a distance image sensor that acquires distance data for a target region;
The object detection system according to claim 1, wherein the control unit causes the distance image sensor to acquire the background data when an object is not detected.   3. The object according to claim 2, wherein the background data is acquired for a target area including a track side of a platform door by the distance image sensor at a timing when the train image is detected by the distance image sensor at the time of the non-detection. Detection system.   The control unit obtains background data for each train entry, and compares the first background data obtained at the time of the current train entry with the second background data obtained at the time of the previous train entry. Item 4. The object detection system according to Item 3.   5. The control unit according to claim 1, wherein the control unit determines that the state is abnormal when a difference value or a detection pixel number difference between the first background data and the second background data exceeds a predetermined limit. The object detection system as described in any one of Claims.   The object detection system according to claim 5, wherein the control unit determines that the state is abnormal when the difference value exceeds the predetermined limit set based on an assumed upper limit value of temperature change.   3. The object detection system according to claim 2, wherein the control unit causes the distance image sensor to acquire the background data when a train line is detected based on an output of the distance image sensor.   The object detection system according to claim 2, wherein the control unit detects an object as a difference between the background data and distance data detected by the distance image sensor at the time of detection.   The said detection part is an object detection system as described in any one of Claims 1-8 which corrects the said background data by performing predetermined offset adjustment with respect to the said background data.   The object detection system according to claim 9, wherein the control unit performs offset adjustment for each pixel or for each block including a plurality of pixels.   The said image sensor is an object detection system as described in any one of Claims 1-10 provided in each of several platform doors provided along the platform end. An object detection method for detecting an object by measuring a target region by an image sensor,
An object detection method of checking an operation state of the image sensor by comparing first background data and second background data acquired at different timings by the image sensor.

Images